Influence of Intensive Diabetes Treatment on Body Weight and Composition of Adults With Type 1 Diabetes in the Diabetes Control and Complications Trial


Diabetes Care. 2001;24(10):1711-172. 

In This Article


Change in Weight During the First Year of Follow-Up

Mean weight and BMI at baseline and changes after 1 year among the adult subjects in each treatment group are shown in Table 1 . Overall, the intensive treatment group gained an average of 3.3 kg compared with 1.2 kg in the conventional group (P < 0.0001). The mean increase in BMI was also significantly higher in the intensive treatment group, both among all subjects (1.2 vs. 0.4 kg/m2, P < 0.0001) and within each sex.

Weight Over the Entire Study Period

The annual distributions of weight and BMI in each treatment group are shown in Fig. 1. From year 1 onward, the medians, the upper and lower quartiles, and even the 5th and 95th percentiles were generally higher in the intensive group. The test of stochastic ordering indicates that these differences attain very high levels of statistical significance over time (P < 0.00001).

Figure 1.

Boxplots of weight (A) and BMI (B) at each annual visit for intensive () and conventional treatment patients (). The central line is the median, the upper and lower boxes represent the upper and lower quartiles, respectively, and the upper and lower vertical lines represent the 95th and 5th percentiles, respectively. The differences between the treatment groups in both measures are significant beginning at year 1 (P < 0.0001). The number of adult subjects in the intensive and conventional treatment groups at each year is noted between A and B.

BMI values in both treatment groups appeared to increase most sharply during the first year and then more gradually thereafter. This raises the question of whether the ongoing increases in BMI eventually level off, or even begin to reverse, in either treatment group. To address this, two-stage random-effects models were used to estimate growth curves characterizing typical changes in BMI over time within each subgroup. Details are given in the APPENDIX. Table 2 summarizes the estimated expected annual rates of change during successive 3-year intervals and contrasts them with the observed average rates over the same periods. Both the expected and observed increases in BMI were uniformly greatest during the first year following randomization, slowing distinctly thereafter. In the conventional treatment group, subsequent rates of increase were only one-half to one-third of the rate during the first year.

Although intensive therapy subjects gained weight less rapidly after the first year, even after 9 years, neither sex demonstrated any tendency to lose the accumulated weight. Among female subjects in particular, any slowing of weight gain was very gradual. If these patterns continued, adult women on intensive therapy would, on average, continue to gain substantial amounts of weight.

Major Weight Gain

The annual prevalence of major weight gain is shown in Fig. 2. The proportion of patients who gained >5 kg/m2 BMI was consistently greater in the intensive treatment group. These differences persisted throughout follow-up, and all were statistically significant at P ≤ 0.01 for both sexes.

Figure 2.

The percentage of adult men (A) and adult women (B) with major weight gain (see RESEARCH DESIGN AND METHODS for definition) in the intensive () and conventional () treatment groups. Differences between treatment groups were significant (P < 0.01) among men at all years after year 1, and at years 2–7 they were inclusive among women. The overall pattern of differences over time was significant at P < 0.01 in both sexes.

Weight Change and Body Composition

BIA was performed once after an average of 70 months of follow-up ( Table 3 ). Among adult women, both mean estimated percent body fat and mean estimated fat-free mass were significantly higher in the intensive treatment group after adjusting for baseline BMI (P < 0.001). Neither parameter differed significantly between treatment groups among men. Overall, 19.5% of the intensive treatment group and 17.0% of the conventional treatment group had increased WHRs (P = 0.284), and this prevalence did not differ significantly between treatment groups within either sex (data not shown). However, among women, waist circumferences were significantly greater in the intensive group (P = 0.0004).

Body composition of patients with and without major weight gain at the time of evaluation (further stratified by sex and treatment group) is described in Table 4 . Wilcoxon tests compare the estimated percentage of body fat and fat-free body mass between treatment groups within each weight-gain stratum and between patients with and without major weight gain in each treatment group. Both characteristics were highly correlated with weight at baseline, so these tests were performed after a linear adjustment for that covariance.

As expected, patients with major weight gain tended to have significantly higher proportions of body fat than similarly treated patients without major weight gain. However, they also tended to have significantly greater fat-free mass. Among patients without major weight gain, intensive therapy was associated with greater fat-free mass but no concomitant difference in adiposity. No treatment-related differences in body composition were observed among patients with major weight gain, but the small number of conventionally treated subjects in these strata limits the statistical power of these comparisons.

The associations among treatment group, weight gain, and body composition were confirmed in simultaneous tests provided by linear analysis-of-covariance models for fat-free mass and percentage of body fat (not shown). In addition to their covariance with baseline weight, both variables showed significant differences associated with sex and major weight gain. Fat-free mass also differed significantly by treatment, whereas percent fat did not; and neither variable suggested any interaction between treatment group and major weight gain. However, fat-free mass did show a nominally significant interaction between treatment and sex.

Effect of Other Baseline and Intrastudy Variables

Baseline variables, including BMI, age, insulin dose, HbA1c, duration of type 1 diabetes, total calories consumed, percentage of calories from fat, and activity level (sedentary compared with not sedentary) were examined to determine their potential impact on weight gain during the study. Intrastudy variables included mean insulin dose, self-reported calorie and fat intake, cigarette smoking, and hypoglycemia. Because of differences between sexes and treatment groups, these analyses were conducted separately in each of the four sex/treatment combinations. Overall, the baseline covariates of age (stratified adjusted relative risk 1.44, CI 1.01-2.05), BMI (1.63, 1.13-2.35), and HbA1c (1.43, 1.00-2.05) were all positively and significantly associated with major weight gain. However, only age was consistent between all four sex/treatment subgroups. None of the intrastudy variables examined was significantly associated with major weight gain.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: